Lane Kenworthy, The Good Society
April 2023

Democracy is a system of decision making in which participants have approximately equal opportunity to influence policy choices.¹ This entails:

1. Each person has the same number of votes in electing policy makers (representative democracy) and in direct policy making (direct democracy).
2. Each person has roughly the same opportunity to influence policy makers’ views and actions via organization, lobbying, protest, and other activities.
3. Individuals have access to adequate information in order to develop informed preferences.
4. Decisions are made according to majority rule (though the majority can’t abridge conditions 1-3).

Democracy is relatively new. As figure 1 shows, two centuries ago hardly anyone lived in a country with a democratic political system. But since then democracy has been on the rise.

The United States has the world’s longest-running political democracy. One of its most remarkable features, highlighted in figure 2, is that for more than two centuries the losing candidates in elections to the presidency, the House of Representatives, and the Senate have peacefully allowed the winning candidates to take office. The government and political system have survived potentially mortal threats, most notably the southern states’ attempted secession and subsequent civil war in the 1860s. Voting rights have been steadily expanded over time — to non-property-owners by the 1830s, to African Americans formally in 1870 and substantively a century later, to women in 1920, to 18-to-20-year-olds in 1971. The system’s stability and popular legitimacy have enabled reasonably wise policy making and helped to facilitate a massive rise in living standards for America’s citizens.

While deserving of praise, America’s democracy is far from perfect. It may discourage voting and other forms of political activity (speech, organization, lobbying, etc.) among Americans with less education and income. The electoral systems for the Senate, the House of Representatives, and the presidency translate votes into representation less effectively than we might wish. Private money may give businesses and wealthy individuals disproportionate influence on policy making. Our veto-point-heavy government structure may go too far in preserving policy continuity at the expense of the majority’s will.

To what extent do these problems exist? Have they worsened over time?

Skip to:

• Unequal voting
• Flawed translation of votes into elected representatives
• Information: from too little to the wrong kind
• Direct democracy: a double-edged substitute
• Unequal influence on policy making
• Veto points subvert majority rule
• Does our political system create an electoral advantage for one of the parties?
• Summary

UNEQUAL VOTING

Does our voting system result in more high-income voters than low-income voters?

As figure 3 shows, in recent decades only about half of voting-age Americans have tended to vote in presidential-year elections. In “off-year” elections, only a third have. These shares have been lower than in many other rich longstanding-democratic nations.

Figure 4 shows voter turnout by socioeconomic status (SES, a measure combining education and income). The higher the socioeconomic status, the more likely the person is to vote.

The class skew in voter turnout owes partly to preferences, but there are systemic obstacles in the US that also matter. Only two-thirds of eligible Americans are registered to vote, because in many states registration must be done in advance and in person. Voting too usually is carried out in person, and on a work day during working hours. Some voting-age adults are barred from voting: noncitizens, prisoners, ex-felons, persons who can’t produce “proper” identification. These obstacles reduce voting more among Americans with lower eduction and/or income. We know that because the class skew in voting widened in the United States when registration requirements were introduced after 1896 and because the skew is larger in the US than in rich countries that have different registration and voting procedures.²

In recent years, as opportunities to vote early and by mail have expanded, turnout has increased significantly (figure 3).³ We don’t yet know whether, and if so how much, this has reduced the class skew in voting.

FLAWED TRANSLATION OF VOTES INTO ELECTED REPRESENTATIVES

An electoral system that effectively translates votes into representation is needed to ensure that each person’s vote has the same weight. America’s electoral system has several problematic features.

Most rich longstanding-democratic countries have a proportional representation electoral system, in which voting is (solely or mainly) for parties and each party’s share of seats in the legislature corresponds to the share of votes it receives. The United States has winner-take-all elections. Votes are cast for candidates, and the candidate receiving the most votes in a district is the only representative of that district. Winner-take-all systems tend to underrepresent minority views and preferences, as smaller parties, the most likely source of such representation, seldom gain political traction. It’s difficult for them to win elections, so their support fades over time and they disappear. Over the past 150 years, no third party has been able to survive in American politics at the national level.⁴

The Senate is one of the most malapportioned legislative bodies in the world, with Americans who live in low-population states far overrepresented relative to their counterparts in high-population states.⁵ As figure 5 shows, the 22 least-populated states have a total population equal to that of California. The 39 million Americans in those 22 states are represented by 44 senators, while the 39 million Americans in California are represented by two.

In 2017 and 2018, for example, Democratic Party senators represented 55% of the US population, Republican Party senators 45% (if we treat each senator as representing half of the population of her or his state). Yet the Democrats had only 48 of the senators, the Republicans 52.⁶ This type of outcome, with the party that got fewer votes ending up with a majority of Senate seats, has occurred at least seven other times in the past century.⁷

Unlike the Senate, the House of Representatives has districts of (approximately) equal population size. Yet it too fails to effectively translate party vote share into party representation share. That’s plainly true for third parties, which seldom get any representation regardless of whether they get 2% of the vote or 20% of the vote. But it’s even true for the two major parties. In the 2016 House elections, for example, Republican candidates got 49.6% of the popular vote but ended up with 55% of the seats.⁸

One reason for this is the heavy concentration of Democratic voters in large cities and hence within certain House districts. This results in more Democratic votes being “wasted” than Republican votes: more Democratic candidates win by a margin of, say, 80% to 20%, whereas Republican winners are more likely do so by a margin of 55% to 45%. Democrats thereby win fewer seats than if their voters were more spread out across House districts.⁹

It’s also partly a result of the gerrymandering of House districts. Every ten years, following the decennial census, the districts are redrawn to respond to population shifts. In most states, the party that controls the state government at that moment gets to decide how to draw the new district boundaries. Partisan gerrymandering consists of creating a number of districts in which your party will win by a comfortable but not excessive margin and a few districts in which the other party will win with nearly all of the votes. This results in more of the other party’s votes being wasted, allowing your party to win a larger number of districts. Republicans were able to do this in more states than Democrats following the 2010 census.¹⁰

Figure 6 illustrates how gerrymandering can be done. Imagine a state with 50 precincts (or 50 voters), 60% of which are Democratic (blue) and 40% of which are Republican (red). The state has five House districts. Those five districts can be drawn in a variety of different ways, some of which accurately reflect the balance within the state as a whole (second column), others that give the Democrats more seats than they should have (third column), and still others that overrepresent the Republicans (fourth column).

A growing number of states have turned over the redrawing of House districts to independent commissions, and this has reduced gerrymandering following the 2020 Census. In the 2022 House elections, Republicans won 50.6% of the popular vote and ended up with 51.0% of the seats.

As with the Senate and the House of Representatives, there is a major flaw in how votes are translated into representation in US presidential elections. Rather than simply tallying up the popular vote, the presidential winner is decided via the electoral college. The candidate who wins the popular vote in each state gets all of that state’s electoral college votes, regardless of the margin of victory.¹¹ When the country is evenly divided between the two parties, there is a significant possibility that the candidate who gets the most votes will nevertheless lose the electoral college.

That has happened only twice in the modern democratic era (since 1920, when women got the right to vote). But those two occurrences, in 2000 and 2016, were among the six most recent elections. It nearly happened also in 2004 and 2020. Had John Kerry gotten 120,000 more votes in Ohio in 2004, he, like George W. Bush and Donald Trump, would have become president despite losing the popular vote. And had Trump gotten 15,000 more votes in Arizona and Georgia and 25,000 more in Wisconsin in 2020, he would have twice in a row won the electoral college despite losing the popular vote.

INFORMATION: FROM TOO LITTLE TO THE WRONG KIND

The earliest systematic studies of Americans’ political beliefs and preferences, in the 1940s and 1950s, found that most citizens don’t pay close attention to most policy issues and so don’t have an opinion on them. This is one reason why only a modest share vote. And those who do vote, even those who are well informed and politically engaged, tend to develop preferences and cast votes based on their group identity, partisan loyalty, and the perceived condition of the economy, rather than informed reflection on the issues.¹²

This owed partly to lack of information. Circumstances today have changed dramatically. With the internet, most citizens have relatively easy access to more than enough information they need to develop informed preferences about issues, parties, and candidates. Yet recent studies, including a replication of the hugely influential The American Voter fifty years on, have found no improvement in citizens’ political awareness or ideological consistency. Political junkies are better informed than ever before, but the average American isn’t.¹³

One reason is that voter ignorance is a product not only of lack of information but also lack of time and interest. It’s also the case, though, that together with the massive increase in access to information has come a parallel increase in faux information and fake news — “facts” and “stories” that are inaccurate or intentionally misleading.

DIRECT DEMOCRACY: A DOUBLE-EDGED SUBSTITUTE

If voters have trouble forming preferences on multiple issues and matching those preferences to candidates, proponents of direct democracy contend that they can do better on individual issues. In about half of the US states, Americans regularly cast yes or no votes on an assortment of initiatives and referendums, making policy decisions directly rather than indirectly through elected representatives. Most of these procedures were created during the progressive era in the early twentieth century.

It isn’t clear whether this has improved our democracy. Some studies conclude that state initiatives and referendums have tended to produce better policy decisions than the regular legislative process.¹⁴ Others point out that those with lots of money have a greater advantage in swaying voters to pass a ballot initiative than in getting lawmakers to follow their wishes, that citizens frequently vote in favor of populist but shortsighted reforms such as term limits and tax cuts, and that the yes-or-no format of direct democracy weakens the quality of policy outputs.¹⁵

UNEQUAL INFLUENCE ON POLICY MAKING

Voting and the translation of votes into representation are democracy’s core inputs. What about the outputs? To what extent do policy decisions reflect the will of the people — or, if not their will, perhaps their interests?¹⁶

One reason for skepticism is that, as with voting, participation in other political activities varies according to socioeconomic status. Wealthy business leaders serve as close advisors to presidents.¹⁷ Firms and groups that represent them (the Chambers of Commerce, National Association of Manufacturers, industry associations) spend large quantities of money lobbying policy makers. Rich Americans shower campaign contributions on their preferred candidates. The most thorough research on this issue, by Kay Lehman Schlozman, Sidney Verba, and Henry Brady, finds that these anecdotal impressions reflect the broader pattern. “The individuals and organizations that are active in American politics,” they write, “are anything but representative. In particular, those who are not affluent and well educated are less likely to take part politically.”¹⁸ As figure 7 shows, Schlozman and colleagues find that Americans with less education and income are less likely to contact a government official, work with fellow citizens to solve a community problem, make a monetary political contribution, work for a political party or candidate, or attend a protest.

Political parties, advocacy groups, and social movements attempt to represent and promote citizens’ interests and preferences. Indeed, according to a distinguished line of political analysis, from E.E. Schattschneider to Thomas Ferguson and Joel Rogers to Jacob Hacker and Paul Pierson, government policy in the United States is determined mainly by the relative strength of organized interest groups.¹⁹ Does this help to compensate for the inequality in citizens’ political participation? Or does it accentuate that inequality?

Businesses and rich individuals have a lot of money, and both the absolute amount and their share have increased since the late 1970s. They’ve also organized into groups — business organizations, foundations, think tanks, political action committees, lobbying agencies — in order to increase their influence on policy making. During this same period, organizations representing ordinary Americans, most notably labor unions, have weakened. Unlike some other affluent democratic nations, the United States has no institutionalized mechanism whereby organizations representing both the more and less powerful discuss or negotiate important policy issues (sometimes called “corporatism”).²⁰ Do affluent Americans and business firms exert disproportionate influence on policy decisions? Has their influence increased in recent decades?

Figure 8 shows the share of high-, middle-, and low-SES Americans who donate money to a political campaign. Figure 9 shows the share of campaign contributions that come from the top .01% of donors. Figure 10 shows the trend in total campaign contributions over the period for which we have reliable data, beginning in 1998. Figure 11 shows the trend in lobbying expenditures. Figure 12 shows the trend in labor union membership. Together, these data tell us that rich individuals and businesses are indeed the most likely to insert money into the political process, that the quantities they give to political campaigns and to lobbyists are large and have increased in recent decades, and that the organizational strength of the chief group representing less-educated and lower-income Americans has declined significantly.

It also is true that Americans with higher incomes are more likely to get the policy they want than are those with middle and low incomes. In separate analyses, Larry Bartels, Martin Gilens, and Gilens and Benjamin Page have found that policy choices are much more likely to reflect the expressed policy preferences of high-income Americans than of those lower on the income ladder.²¹ Figure 13 depicts this finding graphically. The preferences of Americans near the top of the income distribution are positively correlated with policies adopted by congress and the president (solid line). The preferences of those with low incomes are, by contrast, uncorrelated with actual policy choices (dashed line).

What about the influence of interest groups? Here research conclusions are mixed. Some studies have documented a number of important instances of policy change and non-change in recent decades in which business firms and associations have gotten what they wanted.²² On the other hand, studies of lobbying have tended to find little or no systematic impact.²³

Given the increase in income inequality in the United States since the late 1970s, we might expect the class skew in political participation to have increased in recent decades.²⁴ Figure 7 (above) suggests, however, that that hasn’t happened.²⁵ One reason might be that the rise in income inequality has consisted mainly of growing separation between the top 1% and everyone else, rather than between, say, the middle class and the poor.²⁶

Has the disproportionate influence of the rich and/or the business community increased in recent decades? Studies that document policy wins by business and affluent Americans haven’t tended to find a rise in the frequency of such wins in recent decades.²⁷ Martin Gilens has examined the correlation between income and influence on policy at various points over the past half century. As figure 14 shows, Gilens finds that the gap in influence between Americans with high incomes and those with low incomes was small during the Johnson presidency, larger during the presidencies of Reagan and Clinton, but then smaller during the first six years of George W. Bush’s presidency.

If inequality of influence in fact hasn’t increased, it may be because money has diminishing returns in politics. It’s possible that we reached the saturation point decades ago, so that additional money no longer buys much, or perhaps any, additional influence.²⁸

VETO POINTS SUBVERT MAJORITY RULE

America’s founders structured the country’s government to safeguard against excessive power of the majority and against radical change.²⁹ They did so by creating a large number of “veto points” — points at which proposed policy changes can be blocked:

• A directly-elected executive (president). In most other rich longstanding-democratic countries, the executive (usually the prime minister) is the leader of the largest party in the parliament, and she or he has no independent lawmaking authority.
• Two legislative bodies (the House of Representatives and Senate) with equal power. Other nations usually have just one house in parliament; where there are two, one tends to be largely ceremonial.
• A judiciary with considerable leeway to void legislation by ruling it unconstitutional.
• A highly federalized government, with state and local governments that have substantial decision-making power.

In addition, for the past century the Senate has operated under a self-imposed rule — the filibuster — that allows as few as 41 of its 100 members to block proposed policy changes. This means passing legislation in the Senate often requires a supermajority of 60 votes, rather than a simple majority of 51.³⁰

With so many veto points, it is, arguably, too difficult for the majority to implement its policy preferences. Benjamin Page and Martin Gilens examined 1,791 policy changes proposed between 1981 and 2002. Figure 15 shows that when more Americans favor a proposed change, it is more likely to get adopted. Yet a relatively small share of proposals actually get passed. Even for proposed policy changes favored by virtually everyone, the probability of adoption is barely more than 50%.

The difficulty in getting majority preferences enacted into policy is accentuated where there are two dominant ideologically-cohesive and ideologically-distinct political parties, each supported by roughly half of the electorate, one of which adopts an extreme oppositional stance. In such a context, a veto-point-heavy government structure may cross the line from caution to gridlock.³¹

Consistent with this worry, we’ve seen growing use of the filibuster during the era of party polarization, as figure 16 shows.

So has the federal government descended into legislative paralysis? One measure suggests yes. Figure 17 shows a steady decline in the number of laws passed by congress.

However, GovTrack, which compiles these data on the number of laws passed notes the following: “Since World War II (the earliest we have data), Congress has typically enacted 4-6 million words of new law in each two-year Congress. However, those words have been enacted in fewer but larger bills. Therefore, the generally decreasing number of bills enacted into law does not reflect less legislative work is occurring.”³² David Mayhew has compiled data on the number of “major” laws passed by congress each year. These are shown in figure 18. There is no indication of a decline.

So it may be that despite our extensive veto points and extensive party polarization, there’s been less gridlock than widely assumed.

DOES OUR POLITICAL SYSTEM CREATE AN ELECTORAL ADVANTAGE FOR ONE OF THE PARTIES?

Are there systemic biases in our political system that favor the election prospects of the Republicans or the Democrats? Demographic developments, especially the rising share of Latinos in the population, have tilted toward the Democrats. The same is true of cohort trends: each group of young Americans tends to be more liberal than the previous.³³ Yet Republicans have fared better in elections since 1980 than they did in the previous half century, as figure 19 shows. Is that due to structural features of America’s democracy?

Since at least the 1950s, Americans with lower incomes have been more likely to vote Democratic while those with higher incomes have tended to vote Republican.³⁴ The class skew in voting, with voter turnout lower among those with less education and incomes, has therefore tended to benefit Republicans.

Several features of our electoral system also tend to advantage the Republican Party. One is systemic: winner-take-all systems such as ours are less conducive to electoral success by left parties than proportional representation systems.³⁵

At the moment, the Senate’s overrepresentation of small states favors the Republicans, because there are more conservative small states than liberal ones. Figure 20 offers one way to see this. It shows state population by the partisan leaning of each state. If we consider states in which a party got more than 53% of the two-party vote to be solidly partisan, there are 15 solidly Democratic states and 23 solidly Republican ones. In the solidly Republican states, there is one senator per 2 million people. In the solidly Democratic states, there is one senator per 4 million people.

As noted above, Republicans in state government had more opportunity to gerrymander House districts following the 2010 census than did Democrats, and quite a few took advantage of that opportunity. This gave Republicans a boost in House elections, and in each election during the 2010s Republicans’ share of seats in the House of Representatives has exceeded their share of votes received. But as noted earlier, the growing use of independent commissions for redistricting may have reduced this problem.

The Republican Party has tended to be more committed than the Democrats to advancing the economic well-being of affluent individuals and business firms. Businesses and the rich have gotten a large and growing share of the country’s income since the late 1970s, and they have strengthened their political organization considerably. We might expect this to have produced a flood of money to Republicans. The Supreme Court’s 2010 Citizens United ruling prohibited restrictions on political campaign spending by organizations, such as firms and unions, opening the door even further to expenditures by outside groups on behalf of their preferred candidate or party.

Yet the evidence doesn’t support the notion that a money advantage has been the key reason behind recent Republican electoral success. Figure 21 shows campaign expenditures by and for Democrats and Republicans since 1998 (the earliest year for which reliable data are available). Republicans have not dominated here.

It’s too soon to be able to render an informed judgment on the impact of the Citizens United decision, but the degree to which it altered the legal landscape is sometimes overstated.³⁶ Before the super PACs (political action committees) and 501(c)(4)s that sprang up after Citizens United, individuals and corporations already could make unlimited donations to 527s. The only difference is that the new organizations are less constrained in naming the candidates they favor or oppose in advertisements running during the two months prior to the election.

Even if money totals do come to systematically favor Republicans at some point in the future, it’s unclear how much that will matter. There are diminishing returns to money in influencing election outcomes: when a lot is already being spent, additional amounts are likely to have limited impact.

Moreover, the chief determinant of the outcome of national elections is the health of the economy. Douglas Hibbs and Larry Bartels point out that presidential election outcomes can be predicted fairly well using just a single measure of economic performance: income growth in the months preceding the election.³⁷ This is displayed in figure 22. On the vertical axis is the incumbent-party candidate’s popular vote margin. On the horizontal axis is the growth rate of real disposable personal income per capita in the middle two quarters (April through September) of the election year, adjusted for how long the incumbent party has been in office. This simple model does a very good job of predicting the vote outcome. Other models can predict even more accurately by including additional factors, but measures of economic performance are central in all of them.³⁸

House and Senate elections are more idiosyncratic than presidential elections, yet the condition of the national economy tends to be a good predictor for them too.³⁹ Incumbency also has a big effect. As figure 23 shows, a very large share of congressional incumbents who run for reelection win — 93% in the House and 81% in the Senate, on average, since 1950. That’s partly because they are able to raise more money than challengers, but it’s also because they develop name recognition, they can point to things they’ve accomplished for the district, and they have multiple ways of getting media exposure.⁴⁰

Money plays a role in election outcomes, to be sure, but it appears to be no more important as a source of Republican advantage, and perhaps less so, than our winner-take-all elections, Senate malapportionment, House district gerrymandering, and the class skew in voter turnout.

SUMMARY

While there is much to cheer in America’s political system, particularly its stability over time, there also is quite a bit to lament. Too few Americans vote, and participation in voting is heavily class-skewed, due in part to needlessly stiff registration and voting requirements. Effective translation of votes into elected representatives is impeded by key aspects of our electoral system — winner-take-all elections, Senate malapportionment, House of Representatives district gerrymandering, and the presidential electoral college. High-income Americans have much more influence on policy decisions than do those with lower incomes. And America’s multitude of governmental veto points may obstruct policy change.

However, the news is not all bad. New opportunities to vote early and by mail have increased voter turnout. As more states have turned electoral redistricting for the House over to independent commissions, gerrymandering has decreased. Inequality of political influence doesn’t appear to have increased since the late 1970s even though income and wealth inequality have risen sharply. And despite the potential for our veto-point-heavy political system to be paralyzed in the context of a polarized and evenly balanced electorate, passage of major legislation by the federal government hasn’t decreased.